Federated infrastructure but user sovereignty: a “third way” between federated and P2P networks.
But we don’t need to go the whole P2P hog to get robust user sovereignty. Cryptography lets us separate who stores and routes data from who truly owns it. We can combine the easier infrastructure design of a federated protocol with the full decentralized security of a peer-to-peer protocol.
I’'l give it my best shot.
Moving from Web 2 (like Reddit, Facebook, etc) to Web 3 a big thing people are wanting to see is “user sovereignty”. Essentially treating the online identity as a thing owned and controlled by the user instead of something that lent out by the platform.
Federated services like Lemmy, help mitigate the power a platform has over its users, but allowing for cooperation between different providers (lemmy.ml vs midwest.social for example) so that the user is less stuck to the platform they choose. That said, users identies and profiles are still owned by the platforms in this model, and moving would require making a new profile on another platform.
Confederal services would stand as another step away from web 2 sites, by separating the user management and authentication to another layer from the message relay layer. This would mean that a client would be able to pick and choose the services that it may use (like message storage, avatar hosting, reverse proxing, etc) but could also act in a peer to peer way too, messaging directly from one users app to another.
I have some more learning to do before I could tackle the “How can we build this?” section properally, but I am tracking it as: build a user-friendly layer for creating the distributed namespace onto of something in the ETH blockchain to create the sole truth auth mechanism, but use off-chain methods for the majority of the interactions.
You got me a little closer to understanding that and I appreciate the effort you put in.
Conceptually I have just that last little bit to understand about how would a user control all that you mentioned, as at some point that “user stuff” needs to be “stored” or “handled” at some point and I can’t yet understand where that happens.
I get federation (after a couple of days of deliberation and study, but confederation just escapes me.
If it makes you feel better I think the question of “where that happens” isn’t a a solved one really! Some examples of answers though are on the blockchain (essentially paying to be stored as part of the consensus mechanism’s data), IPFS (either volunteer or personally hosted off of it, or through a paid service to ensure it’s hosted somewhere), or stored locally by the user either manually, as part of a hardware token, or in browser cache.